Create app.py

app.py

@@ -0,0 +1,40 @@
+import streamlit as st
+from PIL import Image
+import numpy as np
+import tensorflow as tf
+from tensorflow import keras
+
+# Load the saved model
+loaded_model = keras.models.load_model('tuned_model_classic.h5')
+
+# Define the class labels (you can customize these according to your problem)
+class_labels = ['Stroke', 'Non-Stroke']
+
+# Streamlit App
+st.title('Image Classifier')
+st.write('Upload an image to classify')
+
+uploaded_image = st.file_uploader("Choose an image...", type=["jpg", "jpeg", "png"])
+
+if uploaded_image is not None:
+    # Read the image and preprocess it
+    image = Image.open(uploaded_image)
+    image = image.convert('RGB')
+    image = image.resize((150, 150))  # Resize to match the model's input shape
+    image = np.array(image)  # Convert PIL image to numpy array
+    image = image / 255.0  # Normalize pixel values (similar to how you did in the model training)
+
+    # Make prediction using the loaded model
+    prediction = loaded_model.predict(np.expand_dims(image, axis=0))[0]
+    predicted_class_index = np.argmax(prediction)
+    predicted_class = class_labels[predicted_class_index]
+    confidence = prediction[predicted_class_index]
+
+    # Display the uploaded image and the prediction
+    st.image(image, caption=f'Uploaded Image', use_column_width=True)
+    
+    # Check if the predicted class is "Non-Stroke" and the confidence is high (you can adjust the threshold)
+    if predicted_class == 'Non-Stroke' and confidence > 0.8:
+        st.write(f'Predicted Class: Uncertain (Possibly both Stroke and Non-Stroke) (Confidence: {confidence:.2f})')
+    else:
+        st.write(f'Predicted Class: {predicted_class} (Confidence: {confidence:.2f})')